OSG Council Monthly Teleconference

• Teleconference Phone Number: 1-510-665-5437
• Meeting ID: 1111
• Date: November 11th, 2008
• Time: 10:00 AM Pacific Time
• Attending: Ruth Pordes (OSG Executive Director), Ann Chervenak (ISI), Rob Quick (OSG Operations), Brad Abbott (DZero), Lothar Bauerdick (US-CMS), Dick Greenwood (DOSAR), Kent Blackburn (OSG Council co-Chair & LIGO)

Agenda/Minutes

Status Report from the OSG Executive Director

• November 7th OSG Executive Director Report to Council

minutes:

Ruth reported that the OSG position statement with respec to EGI is being discussed in several different communities. She is planning to talk to the NSF about this at the SC08 Conference.

Usage is about 25% higher than it was at this time last year.

Regarding the 60% job success reported by nanoHub, most VO are seeing higher rates than this for job success. The other exception is DZero but their issues are understood at this time. A team is in place to work with nanoHub to improve this rate.

SBGrid is very happy with all the support they have been getting from the OSG. They plan to publish a scientific paper in December based on result they have obtained with help from the OSG.

It has become apparent that each new use case and/or application that comes to the OSG posses new problems and challenges. We need to have fast loops for understanding the problems and troubleshooting the issues in these new use cases.

Accounting inaccuracies were reported to be roughly 1%, and characterized as "edge" effects. This is not a huge problem for the big users, but for the smaller scale users this can result in a large discrepancy.

The "Value offered by the OSG" paper is being developed and will be discussed at the December 9th OSG Council meeting. It will also be presented to the Joint NSF/DOE Reviewers

Joint NSF/DOE Review - Dates Now Set:

• January 13th, 2009: rehearsal day
• January 14th, 2009: full day of review
• January 15th, 2009: half day of review

minutes:

Currently working with the JOT to get the charge. Ruth spoke with the NSF members of the JOT last week. The DOE is primarily responsible for laying out the venue for this year's Review and has asked that OSG sho involvement from other SciDAC? groups. This could be complex given the limited time available in the review. Ruth reported that the OSG Executive Team is investigating how best to integrate this.

The NSF is interested in hearing what ATLAS and CMS need from the OSG covering the time from roughly now until the 2014/2015 time frame. Ruth is working with Brian and others to go through the exercise establishing a strawman set of goals and metrics. This will be presented to ATLAS and CMS prior to the Review for feedback. ATLAS was not on the call, but Lothar said that CMS thought this is a good idea.

Ruth reported that the final summary report from last year's Joint NSF/DOE Review was never received. She is working with the agencies to identify a way to assure we get a summary report out of this review.

Due to the emphasis on SciDAC? in this Review, it is expected that the nature will be somewhat different and the "charge given to the reviewers (ie the questions we will plan our presentations to)" for last year's review will not be very relevant to the planning process.

Recent OSG Usage (See Plot Below)

minutes:

The plot shows the same increase in number of jobs as was reported by Ruth earlier. Brian has been working on estimating the total number of cores within the OSG and has a rough estimate of 40-50 thousand cores within OSG. The claim is that this gives us plenty of head room. OSG is looking to have a filler application, hopefully a science application that can fill in all the unused job slots. Ruth said there were a couple of candidates for this. It was also noted that it would be important to have a throttle-back mechanism available to the sites. It was also reported that Engage has already seen that when they attempt to increase the number of jobs running, they start to see a queue.

OSG All-Hands-Meeting Program Committee and Strawman Agenda

• We proposed:

1. half day of "State of the OSG"
2. half day of "Site Administration"
3. half day of "Atlas/CMS Workshops"
4. two half days of Panels
   1. Panel on Virtual Machine Technology and its applications to OSG
   2. Panel on Community Grids and the implications for OSG
5. half day of LIGO
   1. "The Science of LIGO"
   2. "The LIGO Data Grid"
   3. "The LIGO Authentication Project"
   4. "LIGO Education Out Reach"
6. Tours of the LIGO Livingston Observatory in groups of ~30 using registration to group people.
   1. I would suggest one tour a day on each day (Mon, Tues, Weds)
7. There was a suggestion of possibly having a Council Dinner Wednesday Night
8. The conference dinner to be held on Tuesday night

All of this is very sketchy at this point. The program committee will dig into the details once we have your inputs.

• Received one recommendation to have Poster Session(s).

minutes:

Speaking to the idea of having an award for best poster, Ann reported that when she has been involved with poster judging, a small committee was formed that spent some time reviewing posters and then gathered together as a group to determine the winner. Lothar reported that at the last LHC Users Meeting, each attendee was given a ballot to cast their individual votes. The votes were placed in a box with someone on guard to assure voters were informed and the winner was selected by the voter rankings. Ann said that at SC Poster Sessions the prize was a gift certificate worth some amount that was good at the SC Store (where coffee mugs, shirts, etc. were available). Kent believes that LIGO Livingston Observatory (LLO) has a similar offering.

Kent asked if CMS and ATLAS were interested in having a half day attached workshop for their respective groups as was the case at RENCI last March. Lothar said this sounded good to CMS. ATLAS was not on line but Kent sent an email to Michael Ernst regarding this.

The members of the council that were online were asked if they plan to be at the All-Hands Meeting. Everyone on the call expected to attend except Brad Abbott (teaching obligation). Lothar noted that he had a conflicting meeting and would only be there for part of the meeting. The council also liked the idea of having a Council Dinner Wednesday night (before the council face-to-face meeting on Thursday).

Other suggestions were:

• Ruth wanted to have a way to bring more small sites to the site admin workshop
• Ruth also wanted to have a half day workshop for VO Managers and Security contacts

Round-Table Virtual Organization Reports and Issues - Ongoing Goals/Milestones

minutes:

• LIGO: Several changes have been made to the configuration of job submission and execution in the last week for Einstein@Home jobs running on the OSG to all better throughput and accounting. OSG is no longer grouped with DGRID in the scientific credits accounting. Also changes were made to accelerate the completion of credit contributions for the OSG.

AOB

• Perspectives on Distributed Computing: Thirty People, Four User Types, and the Distributed Computing User Experience

minutes:

Ruth reported that the content of this report is likely to be considered in the Joint DOE/NSF Review in January. Ann reported that Lisa is very interested in hearing what OSG's thoughts are regarding this report. Ruth is planning to have the document reviewed for OSG content. The Council decided that it would be good for Lisa Childers (one of the authors) to present on this at the December 9th OSG Council Telecon. Kent will put this on the agenda and ask the Council to review the document before next month's meeting. Ann will contact Lisa to ask her to put our Council Meeting on her agenda.

-- KentBlackburn - 05 Nov 2008